Upload
others
View
4
Download
0
Embed Size (px)
Citation preview
MinPET – 4IR, Tech Transfer, Innovation, Commercialisation
4 Dec 2020 Simon Connell : Tastes of Nuclear Physics 1
Diamonds are rare but valuable• 1.7 cpht ~ 3.4 parts per billion.• In “lumps”, not finely dispersed.• Extreme case for high value buried in waste• Global rough diamond market 153 M ct, 2017 ~ 15.4 B$• Breakage also significant driver
MinPET Benefits• Less plant, less processing less energy, less water, greener• Marginal mines viable, longer life of mine, mine in arid regions• Audit tailings and return to greenfield • Protect diamonds from breakage
4IR
Diamond History
4 Dec 2020 Simon Connell : Tastes of Nuclear Physics 2MinPET at UP 224-08-2017
• Natural diamond genesis was mainly billions of years ago, a much younger earth.
• They had their genesis at 200-600 km depth, and then along residence at about 100km depth
• They were brought to the surface tens of millions of years ago, transported by volcanic material.
• They are now either in kimberlite pipes or dispersed alluvially.
• There won’t be any new natural diamond production.
• We need to protect what we have.
Diamond History
4 Dec 2020 Simon Connell : Tastes of Nuclear Physics 3
Wikipedia
Earliest use of diamond by humans4500 BC
Chinese axes
Archeological evidence that the Chinese could polish Corrundum, implying they had some diamond
Increasing hardness
Diamond History
4 Dec 2020 Simon Connell : Tastes of Nuclear Physics 4
The Valley of the Diamonds
Diamond mining in the Indian Kingdom of Motupalli
1600’s
From “The Hope Diamond”by Richard Kurin(not original ref)
MinPET – 4IR, Tech Transfer, Innovation, Commercialisation
4 Dec 2020 Simon Connell : Tastes of Nuclear Physics 5
MinPET developments have benefitted from ATLAS at CERN
• Sensor development• Simulation• High data throughput• Big Data analysis• High Performance Computing• Human resources• Machine learning
Dark Matter
What is PET ……. Medical PET …… the idea
4 Dec 2020 Simon Connell : Tastes of Nuclear Physics 6
Positron EmmissionTomography
positron
e-
Beta decay
Inject patient with radio-labelled sugar / biomolecule.Targets a metabolic function.• Cancer metabolizes fast• Active areas of the brain
too
What is PET ……. Mineral PET …… the idea
4 Dec 2020 Simon Connell : Tastes of Nuclear Physics 7
Positron EmmissionTomographypositron
e-
Beta decay
Inject patient with radio-labelled sugar / biomolecule.Targets a metabolic function.• Cancer metabolizes fast• Active areas of the brain
too
Energy (MeV)20 25 30 35 40
Cro
ss-s
ectio
n (m
b)
01234567
12C(g,n)11C
The extreme example of something very valuable in barren rock
4 Dec 2020 Simon Connell : Tastes of Nuclear Physics 8
Mirny Diamond Mine in Siberia, Russia, very big ….
But now mined out !
A lot of rock must be processed !
Where are diamonds to be found ?
4 Dec 2020 Simon Connell : Tastes of Nuclear Physics 9
Revenue(M$pa)
Produc2on(Caratspa)50000 168000000
53LargestDiamondMines2015127-million carats 100 million tons$14 billion
• 6000 known kimberlite pipes.
• 1% commercial• 60 diamond
mines
Tradiational diamond recovery is expensive, destructive
4 Dec 2020 Simon Connell : Tastes of Nuclear Physics 10
Successive stages of crushingCentrifuge
Greasetable
Expensive à breakage
The Diamond breakage issue
4 Dec 2020 Simon Connell : Tastes of Nuclear Physics 12
51.27 carats19.46 carats
Diamond Breakage – Loss of Revenue
US$ 215 000US$ 6 900 000
Total value as two stones: US$ 7 135 000Total value as one unbroken stone: US$ 17 700 000
US$ values are approximateEstimate loss of $10m
The Diamond breakage issue
4 Dec 2020 Simon Connell : Tastes of Nuclear Physics 13
The Cullinan, about 3107 ct, was maybe twice as big ?
Stippled face vs cleaved face ….. breakage ?
MinPET in a Nutshell
4 Dec 2020 Simon Connell : Tastes of Nuclear Physics 14
planar PET array
ejection system
irradiation hopperproduce
PET isotopes
hold hopper ~30 min 11C dominates
Activation System
Detection System
Image ReconstructionClassification
30 MINUTE HOLD
HOPPER
PET Lines of Response (LoR)
4 Dec 2020 Simon Connell : Tastes of Nuclear Physics 15
positron
e-
11C beta decay
Irradiation converts normal 12C to 11C by irradiation with high energy photons
…. Detection ………. Reconstruction ……
PET Lines of Response (LoR)
4 Dec 2020 Simon Connell : Tastes of Nuclear Physics 16
Reconstruction of a Point SourceBack projection
Need1,000,000 Lines of Response per rock.
Technological developments compensated unforeseen systematic problems exactly.
Comparison of activated and normal kimberlite
4 Dec 2020 Simon Connell : Tastes of Nuclear Physics ‹#›
Activated kimberlite vs normal kimberlitesingles spectra (after 60min hold time)
• No long lived species
• PET signal is dominant
17
MinPET Activation of kimberlite is very transitory
MinPET in a Hospital
4 Dec 2020 Simon Connell : Tastes of Nuclear Physics 18
First “killer” experiment (2007) Karolinska Hospital
Medical Gamma therapy + Medical PET
5cm
“Patient”
Detector development
4 Dec 2020 Simon Connell : Tastes of Nuclear Physics 19
Detector Development – with Italian Partner
BGO segmented scintillator crystals
Light guide concentrator to SiPM
High throughput front end electronics
Housing
R&D jointly Italy and SA
Detector development
4 Dec 2020 Simon Connell : Tastes of Nuclear Physics 20
• IAEA Country Programme Framework Technical Cooperation: circa R2.8M • Development of Mk3 detectors with NeT Instruments• TIA seed fund: visit to commission detectors
Typical results
4 Dec 2020 Simon Connell : Tastes of Nuclear Physics 21
10 cm cube kimberlite2.3kg
12.5 carat diamond
Typical results
4 Dec 2020 Simon Connell : Tastes of Nuclear Physics 24
10 cm cube kimberlite2.3kg
7 mm diamond2.9 carat
Typical results
4 Dec 2020 Simon Connell : Tastes of Nuclear Physics 25
CaCO340Ca -> 38K16O -> 15O12C -> 11C
All three are PET isotopesWorst case of “background”
Typical results
4 Dec 2020 Simon Connell : Tastes of Nuclear Physics 26
1.431 kg of CalciteL = 15cm
10.3 ct diamond
Typical results
4 Dec 2020 Simon Connell : Tastes of Nuclear Physics 28
350g kimberlite (50mm)
2.9 carat diamond
Typical results
4 Dec 2020 Simon Connell : Tastes of Nuclear Physics 29
350g kimberlite (50mm)
0.23 carat diamond4 x 5.5 x 0.6 mm3
F = 3mm
Typical results
4 Dec 2020 Simon Connell : Tastes of Nuclear Physics 30
410g drill coref = 74 mmL = 40 mm
5.0 carat diamond
Typical results
4 Dec 2020 Simon Connell : Tastes of Nuclear Physics 31
3.33kg kimberlite
22.9 carat diamond
18 cm
10 cm
Full Physics Sim : Geant4 Monte Carlo
4 Dec 2020 Simon Connell : Tastes of Nuclear Physics 34
Full physics model of MinPET
• Electron accelerator • Mixed radiation field from the Activation stage
• Treat Nuclear Physics with TENDL • GDR X(g,n)Y• Other channels and 2nd order processes
• Simulate moving rockstream• Each time slice separate geant4 run
• Simulate Radiation Damage in Diamond• Simulate Activation in Kimberlite• Simulate Detector Response• Produce Data nTuples (Sim or Exp)• Create 4D Sinogram• Perform 3D Reconstruction (MLE - GPU)• Perform AI Classification
Full Physics Sim : Diamond Damage by MinPET
4 Dec 2020 Simon Connell : Tastes of Nuclear Physics 35
Estimate vacancies produced by shower and secondaries from 40 MeV e-g beamGR1 @ 0.3 ppb
Paleo-exposure to diamond
Paleo-dose10 Bq/g integrated back in time for 2 billion years
Estimate primary vacancies produced by natural photons > 185 keV
Nature = 104 x MinPET dose (+ some annealing of V and aggregation of defects)
MinPET’s effect on diamonds is many orders less than exposure in nature …Damage study concludes beow low temp laser induced Photo Luminescence
Artificial Intelligence, Big Data, High Performance Computing
4 Dec 2020 Simon Connell : Tastes of Nuclear Physics 36
We now have an “all singing and all dancing”Artificial Intelligence to classify the diamonds.
Detection à Classificationt < 5 s N
umbe
r of c
lass
ifica
tions
~ significance
Training data from Simulation• Big effort to develop accuracy in SIM• Benchmarking to experiment• High performance computing• Large data sets
Full Physics Sim : Validation Scale-up tool
4 Dec 2020 Simon Connell : Tastes of Nuclear Physics 37
Take home message .....
Demonstrated MinPET in a FULL-DRESS-REHEARSALRate equivalent to 1 MinPET unit ... Run of Mine ... 500 tph
Benchmarked Computer Simulations for FDR
Complete exclusion of “unknown unknowns” to 10%
Validatedscale-up tool
A FDH is when all actors
perform with all costumes
and props in sequence
High Throughput data
4 Dec 2020 Simon Connell : Tastes of Nuclear Physics 38
https://cds.cern.ch/record/2244345/files/ATL-DAQ-PROC-2017-007.pdf
16 GB/s
20 x 1 GB/s ….......
1 GB/s stored for 25s
ATLAS MinPET
Randell Heyman
Scale-up : Data Rates - we need to process 16 GBs How does ATLAS do this with 160 GBs ?
Partition data into Regions of Interest, analyze in parallel, then data fusion
Gem Diamonds own Letśeng mine in Lesotho
4 Dec 2020 Simon Connell : Tastes of Nuclear Physics 39
www.gemdiamonds.com
The Lesotho Legend
910 carats
Lesotho Promise603 carats
Letšeng Star550 carats
Letšeng Legacy493 carats
Leseli La Letšeng478 carats
Gem DiamondsAnnual Report and Accounts2017
Joint research with Gem Diamonds
Producer of especially large
PolyPET – Element sensitive PET
4 Dec 2020 Simon Conneall : Tastes of Nuclear Physics 41
Analysis of• Coal C – Calorific Value O – Moisture
• Explosives• Narcotics• Plastics
PolyPET – Element sensitive PET
4 Dec 2020 Simon Connell : Tastes of Nuclear Physics 42
• Carbon as a proxy for Calorific value• Oxygen as a proxy for Moisture content
Must be on-line, real-time
Measure Coal quality
PolyPET – Element sensitive PET
4 Dec 2020 Simon Connell : Tastes of Nuclear Physics 43
hHPGe_0000Entries 32768Mean 516.7Std Dev 6.932
490 500 510 520 530 5400
1000
2000
3000
4000
5000
6000
hHPGe_0000Entries 32768Mean 516.7Std Dev 6.932
hHPGe_0000Entries 32768Mean 513.4Std Dev 14.87
hHPGe_0001Entries 32768Mean 513.3Std Dev 15.26
hHPGe_0002Entries 32768Mean 513.1Std Dev 15.65
hHPGe_0003Entries 32768Mean 513Std Dev 16.19
hHPGe_0004Entries 32768Mean 512.9Std Dev 16.91
hHPGe_0005Entries 32768Mean 512.5Std Dev 17.93
hHPGe_0006Entries 32768Mean 512.1Std Dev 19.09
hHPGe_0007Entries 32768Mean 511.7Std Dev 20.25
hHPGe_0008Entries 32768Mean 511.3Std Dev 21.35
hHPGe_0009Entries 32768Mean 510.7Std Dev 22.37
hHPGe_0010Entries 32768Mean 510.4Std Dev 23.28
hHPGe_0011Entries 32768Mean 510.1Std Dev 23.81
HPGe tA=257.93s
0 100 200 300 400 500 600 700Time [minutes]
3-10
2-10
1-10
1
10
210
310
410
dNg/
dT DataBackgroundFit CombinedFit Oxygen-15Fit Carbon-11Fit Potassium-38 and Iron-53Fit long
• Time differential detection of PET activity• Each isotope has its own half life• Elemental analysis by lifetime analysis
PolyPET – Element sensitive PET
4 Dec 2020 Simon Connell : Tastes of Nuclear Physics 44
PolyPET sampler3kg sample every 5 min
Main supply belt
Activation40 MeV electrons
Bremsstrahlung target
PET Scanner5 min sampling
Total Body PET
4 Dec 2020 Simon Connell : Tastes of Nuclear Physics 45
• 40 x more efficient• Lower dose, multiple scans• Time variation (mili seconds) – PET Video• Time variation (minutes) - Dynamics
Disruptive Medical Technology.
Innovation, Commercialisation
4 Dec 2020 Simon Connell : Tastes of Nuclear Physics 46
1. Fundamental Research + Idea2. Disclosure then Protection of IP with TTO
• 11 Patents related to MinPET• Secrecy
2. Market Research3. Economic Modelling4. Business Case5. Plant Integration6. Benchmark Competitor Technologies7. Fund Raising8. Proto-type Development9. Technology Readiness Levels10. Scale – up11. Legal Issues (challenges to IP)12. Various impediments (need grit)13. Discussion to levels of CEOs in companies14. Venture capital15. Technology Partners16. Customers
Students Welcome
https://panip.febe.uj.ac.za
Partners, Funders
4 Dec 2020 Simon Connell : Tastes of Nuclear Physics 47
Engineering Company Conversations
Research collaborations :
Funders :
These examples show the approved forms of application for the primary horizontal logo. The full-colour logo may only be applied to a white background. The logo may only be used as a one-colour application in white, black and UJ orange.
Please note that the one-colour logo may ONLY be applied on a white, black or UJ orange background and NOT on a faculty colour background.
The colour usage may not deviate from the examples given, unless otherwise specifi ed in this manual.
180º
270º
225º
135º
CORPORATE COLOUR USAGE
UJ LOGOS | CORRECT LOGO USAGE | CORPORATE COLOURS | PRIMARY HORIZONTAL LOGO
Full colour
One colour
IP jointly held by :These examples show the approved forms of application for the primary horizontal logo. The full-colour logo may only be applied to a white background. The logo may only be used as a one-colour application in white, black and UJ orange.
Please note that the one-colour logo may ONLY be applied on a white, black or UJ orange background and NOT on a faculty colour background.
The colour usage may not deviate from the examples given, unless otherwise specifi ed in this manual.
180º
270º
225º
135º
CORPORATE COLOUR USAGE
UJ LOGOS | CORRECT LOGO USAGE | CORPORATE COLOURS | PRIMARY HORIZONTAL LOGO
Full colour
One colour
These examples show the approved forms of application for the primary horizontal logo. The full-colour logo may only be applied to a white background. The logo may only be used as a one-colour application in white, black and UJ orange.
Please note that the one-colour logo may ONLY be applied on a white, black or UJ orange background and NOT on a faculty colour background.
The colour usage may not deviate from the examples given, unless otherwise specifi ed in this manual.
180º
270º
225º
135º
CORPORATE COLOUR USAGE
UJ LOGOS | CORRECT LOGO USAGE | CORPORATE COLOURS | PRIMARY HORIZONTAL LOGO
Full colour
One colour
For Nico, and all of you …. Thank YouThe warriors of the Zulu Impi called out to the young men: Grow up to be yet braver than we !